Non-small cell lung cancer (NSCLC) patients with epidermal growth factor receptor (EGFR)-activating mutations respond well to EGFR-targeted tyrosine kinase inhibitors (TKIs). However, within a year, they generally relapse. Improving the long-term outlook of TKI treatment is a compelling challenge. I isolated seven single-cell derived clones from the human lung cancer cell line PC9, which contains an activating EGFR mutation and is an established model for erlotinib-responsive NSCLC. The short-term erlotinib response of these PC9 clones, measured by GI50, was comparable to PC9 parental. In contrast, I detected variability among clones in the long-term response, measured by imaging the cells at high time resolution during a prolonged drug exposure. I then categorized the clones as having low, medium, and high erlotinib sensitivity. These clones serve as a model to study the influence of variable initial conditions of TKI sensitivity onto overall treatment efficacy in EGFR-mutant cancers. This variability has not been previously characterized or studied, to my knowledge. I hypothesize that this TKI response heterogeneity is responsible, at least in part, for early- stage TKI treatment failure and/or the emergence of resistant cell populations in the long term. To test this hypothesis, I propose an integrated approach of modeling and experimentation, articulated in 3 aims. In Aim1, I plan to characterize cellular metrics that quantify the differential erlotini responses in each PC9 clone. To this end, rates of proliferation, death and cycle arrest will be assessed by extended live-cell fluorescence microscopy assays, under varying conditions of erlotinib exposure. Differential rates will shed light on the cellular dynamics that underlie variability of erlotinib response. In Aim2, I will extend these dynamics to the time scale (months) of in vivo responses, by constructing a mathematical compartment model that predicts long- term outcomes, which incorporates measured rates. In Aim3, I will test whether, under erlotinib treatment, activation of RTKs other than EGFR sustain downstream signaling and survival in low-sensitivity PC9 cells.I will answer these questions using protein arrays and quantitative western blotting. Each of the aims will provide novel information on the dynamics of erlotinib response in the PC9 clone model by quantifying its heterogeneity, probing its possible molecular basis, and modeling long-term outcomes. This integration of cellular, molecular and computational data may provide critical insights into the dynamics of erlotinib short- and long-term failure. This knowledge would help reconcile the promise of targeted therapies with the challenge of tumor heterogeneity, and could have high translational value.